Difference between revisions of "Recall Bias"

Jump to: navigation, search
(2 intermediate revisions by one other user not shown)
Line 22: Line 22:


== Additional Resources ==
== Additional Resources ==
*DIME Analytics’ guidelines on [https://github.com/worldbank/DIME-Resources/blob/master/survey-instruments.pdf survey design and pilot]
* [https://www.ifpri.org/blog/do-you-remember-measuring-anchoring-bias-recall-data Measuring anchoring bias in recall data] from the IFPRI Research Blog
* Development Impact Blogpost on [http://blogs.worldbank.org/impactevaluations/decomposing-response-error-improve-consumption-survey-design Response Error in Consumption Surveys] and the related [http://documents.worldbank.org/curated/en/122481467999693721/pdf/WPS7646.pdf paper]
* Development Impact Blogpost on [http://blogs.worldbank.org/impactevaluations/decomposing-response-error-improve-consumption-survey-design Response Error in Consumption Surveys] and the related [http://documents.worldbank.org/curated/en/122481467999693721/pdf/WPS7646.pdf paper]
* Blog from Financial Access on [http://www.financialaccess.org/blog/2015/7/30/reliability-of-self-reported-data-recall-bias The Reliability of Self-reported Data]
* Blog from Financial Access on [http://www.financialaccess.org/blog/2015/7/30/reliability-of-self-reported-data-recall-bias The Reliability of Self-reported Data]
Line 27: Line 29:
** Abstract: Between 2000 and 2002, we followed 1621 individuals in Delhi, India using a combination of weekly and monthly-recall health questionnaires. In 2008, we augmented these data with another 8 weeks of surveys during which households were experimentally allocated to surveys with different recall periods in the second half of the survey. We show that the length of the recall period had a large impact on reported morbidity, doctor visits; time spent sick; whether at least one day of work/school was lost due to sickness and; the reported use of self-medication. The effects are more pronounced among the poor than the rich. In one example, differential recall effects across income groups reverse the sign of the gradient between doctor visits and per-capita expenditures such that the poor use health care providers more than the rich in the weekly recall surveys but less in monthly recall surveys. We hypothesize that illnesses – especially among the poor – are no longer perceived as “extraordinary events” but have become part of “normal” life. We discuss the implications of these results for health survey methodology, and the economic interpretation of sickness in poor populations.
** Abstract: Between 2000 and 2002, we followed 1621 individuals in Delhi, India using a combination of weekly and monthly-recall health questionnaires. In 2008, we augmented these data with another 8 weeks of surveys during which households were experimentally allocated to surveys with different recall periods in the second half of the survey. We show that the length of the recall period had a large impact on reported morbidity, doctor visits; time spent sick; whether at least one day of work/school was lost due to sickness and; the reported use of self-medication. The effects are more pronounced among the poor than the rich. In one example, differential recall effects across income groups reverse the sign of the gradient between doctor visits and per-capita expenditures such that the poor use health care providers more than the rich in the weekly recall surveys but less in monthly recall surveys. We hypothesize that illnesses – especially among the poor – are no longer perceived as “extraordinary events” but have become part of “normal” life. We discuss the implications of these results for health survey methodology, and the economic interpretation of sickness in poor populations.
* Kathleen Beegle, Calogero Carletto, Kristen Himelein, [http://www.sciencedirect.com/science/article/pii/S0304387811000939 Reliability of recall in agricultural data], In Journal of Development Economics, Volume 98, Issue 1, 2012, Pages 34-41, ISSN 0304-3878, https://doi.org/10.1016/j.jdeveco.2011.09.005.
* Kathleen Beegle, Calogero Carletto, Kristen Himelein, [http://www.sciencedirect.com/science/article/pii/S0304387811000939 Reliability of recall in agricultural data], In Journal of Development Economics, Volume 98, Issue 1, 2012, Pages 34-41, ISSN 0304-3878, https://doi.org/10.1016/j.jdeveco.2011.09.005.
** Abstract: Despite the importance of agriculture to economic development, and a vast accompanying literature on the subject, little research has been done on data quality. Due to survey logistics, agricultural data are usually collected by asking respondents to recall the details of events occurring during past agricultural seasons, potentially leading to recall bias. The problem is further complicated when interviews are conducted over the course of several months, thus leading to recall of variable length. To test for recall bias, the length of time between harvest and interview is examined for three African countries with respect to several common agricultural input and harvest measures. The analysis shows little evidence of large recall bias impacting data quality. There is some indication that more salient events are less subject to recall decay. Overall, the results allay some concerns about the quality of some types of agricultural data collected through recall over lengthy periods.
** Abstract: Despite the importance of agriculture to economic development, and a vast accompanying literature on the subject, little research has been done on data quality. Due to survey logistics, agricultural data are usually collected by asking respondents to recall the details of events occurring during past agricultural seasons, potentially leading to recall bias. The problem is further complicated when interviews are conducted over the course of several months, thus leading to recall of variable length. To test for recall bias, the length of time between harvest and interview is examined for three African countries with respect to several common agricultural input and harvest measures. The analysis shows little evidence of large recall bias impacting data quality. There is some indication that more salient events are less subject to recall decay. Overall, the results allay some concerns about the quality of some types of agricultural data collected through recall over lengthy periods.


[[Category: Questionnaire Design]]
[[Category: Questionnaire Design]]

Revision as of 15:02, 12 June 2019

Recall bias is bias caused by inaccurate or incomplete recollection of events by the respondent. It is a particular concern for retrospective survey questions.

Read First

  • Research shows that lower salience and longer recall periods increase forgetfulness [1] [2]

Guidelines

How long is "too long" for recall?

It depends on the type of event respondents are being asked to recall. Research shows strong evidence of recall bias in food consumption, but little evidence for agricultural production. As a rule of thumb, infrequent events (e.g. purchases of major assets) will be memorable for longer periods of time than routine events (e.g. use of public transportation).

How to avoid recall bias?

Useful strategies:

  1. Reduce recall periods as much as possible. For example, add follow-up surveys by phone, or personal diaries.
  2. Conduct focus groups to understand salience of the indicator in question, and gauge a reasonable recall period.
  3. When Piloting your Survey, carefully test recall periods; if possible try shorter and longer periods and check for differences in variance


Back to Parent

This article is part of the topic Questionnaire Design


Additional Resources

  • DIME Analytics’ guidelines on survey design and pilot
  • Measuring anchoring bias in recall data from the IFPRI Research Blog
  • Development Impact Blogpost on Response Error in Consumption Surveys and the related paper
  • Blog from Financial Access on The Reliability of Self-reported Data
  • Jishnu Das, Jeffrey Hammer, Carolina Sánchez-Paramo, The impact of recall periods on reported morbidity and health seeking behavior, In Journal of Development Economics, Volume 98, Issue 1, 2012, Pages 76-88, ISSN 0304-3878
    • Abstract: Between 2000 and 2002, we followed 1621 individuals in Delhi, India using a combination of weekly and monthly-recall health questionnaires. In 2008, we augmented these data with another 8 weeks of surveys during which households were experimentally allocated to surveys with different recall periods in the second half of the survey. We show that the length of the recall period had a large impact on reported morbidity, doctor visits; time spent sick; whether at least one day of work/school was lost due to sickness and; the reported use of self-medication. The effects are more pronounced among the poor than the rich. In one example, differential recall effects across income groups reverse the sign of the gradient between doctor visits and per-capita expenditures such that the poor use health care providers more than the rich in the weekly recall surveys but less in monthly recall surveys. We hypothesize that illnesses – especially among the poor – are no longer perceived as “extraordinary events” but have become part of “normal” life. We discuss the implications of these results for health survey methodology, and the economic interpretation of sickness in poor populations.
  • Kathleen Beegle, Calogero Carletto, Kristen Himelein, Reliability of recall in agricultural data, In Journal of Development Economics, Volume 98, Issue 1, 2012, Pages 34-41, ISSN 0304-3878, https://doi.org/10.1016/j.jdeveco.2011.09.005.
    • Abstract: Despite the importance of agriculture to economic development, and a vast accompanying literature on the subject, little research has been done on data quality. Due to survey logistics, agricultural data are usually collected by asking respondents to recall the details of events occurring during past agricultural seasons, potentially leading to recall bias. The problem is further complicated when interviews are conducted over the course of several months, thus leading to recall of variable length. To test for recall bias, the length of time between harvest and interview is examined for three African countries with respect to several common agricultural input and harvest measures. The analysis shows little evidence of large recall bias impacting data quality. There is some indication that more salient events are less subject to recall decay. Overall, the results allay some concerns about the quality of some types of agricultural data collected through recall over lengthy periods.